1During the 1990s and 2000s, in secularized societies also concerned by upheavals associated with globalization, an increase in religious practice and above all religious attitudes can be observed among the youngest age groups of immigrants and descendants of immigrants originating from the Maghreb, the Sahel and to a lesser degree Turkey. In recent years, religious practices in the broadest sense have become increasingly pronounced in France both among immigrants and descendants of immigrants from Muslim countries. A similar observation has been made in the Netherlands among immigrants and descendants of immigrants from Morocco and Turkey (Maliepaard and Gijsberts 2012). This development of practices and attitudes is specific to immigrants from the South, but there are also some discernible signs among natives of neighbourhoods with a large immigrant presence. The growth in religiosity of those who declare themselves to be Muslim began in the mid-1990s and outlines a tendency that runs counter to the development of Europe as a whole.
2From where does this increase in religiosity among immigrants from the South in strongly secularized countries stem? Why is it more evident in age groups that in Europe are ordinarily the most indifferent to beliefs and the most uninvolved in religious practices? The aim of this analysis is firstly to understand the reasons for this religious revival, which has been reported by the press and many observers in European countries. What does it owe to outside influences and to living conditions in the host society? How can we understand the behaviour and attitudes of individuals who immigrated when they were children and who are thus partially socialized in Europe? What could be the factors that incite young members of these groups to distance themselves from the behaviour that prevails in their host country?
3Certainly, the problems do not arise for first generation immigrants in the same terms as they do for those of the second generation. In contrast to their parents, descendants of immigrants did not immigrate and the issue of identity is undoubtedly more open and more sensitive for them since they are in a position to define their identity through borrowings and rejections specific to two spaces: the host society and their society of origin. It is recognized that practices and attitudes and thus religious practices and attitudes are formed in part during socialization within primary groups—family, neighbourhood friends and peer groups. We know that determinants that arise from heritage are diluted in passing from one generation to another, from parents to children, and that in contrast those forged in response to the needs of the present (life’s uncertainties, desire for self-assertion or for recognition) can diminish or conversely increase with social anxiety and relations with peers. Aging is normally accompanied by an increase in religiosity, notably because the issues that call for a religious response become more acute with age. In societies undergoing secularization, new generations feel a lesser presence of religious values in comparison to those that preceded them. Without being able to encompass the full spectrum of determinants mentioned, we first hypothesize here that religious vitality among immigrants and descendants of immigrants from the South is stimulated by material difficulties and a weakening of the guarantees provided by secular social protection. If we imagine that the resurgence in religiosity among immigrants and above all their descendants is linked to their living conditions in Europe, specifically relative level of education, social insecurity and segregation, we cannot set aside the idea that the attitudes and behaviours observed reflect tension born from contacts between the different reference spaces. The approach adopted is based on the hypothesis that the moral context during the transmission of religious practices and attitudes is important. A comprehensive schema must consider the relationship between the respective developments of the host country and country of origin as well as local norms relating to the context of residence, taking into account the degree of segregation and community closedness of neighbourhood of residence and finally familial norms (Van Tubergen 2006). This leads us to consider members of migrant ethnocultural minorities to be actors provided with resources, albeit limited: their behaviour and attitudes contribute to the realization of life goals relating to their trajectories and a shared identity.
4Assimilation presupposes a certain unity of the normative and cultural models of a host society where, through the force of circumstances, descendants of immigrants and even immigrants themselves, tend to depart from the attitudes and behaviours characteristic of their countries of origin to adopt those of their host society to various degrees (Alba and Nee 2003; Waters and Jimenez 2005). The acculturation process, even when not aimed at assimilation, implies a partial abandonment of origins.  Inherited habits erode—language is lost, customs and religion are overcome by another credo—, ways of doing things approach those of the host county or at least intermingle. Influences built up over the course of decades interfere with more cyclical determinants. For immigrants and descendants of immigrants, the values and norms of host societies function either as an attractor or sometimes as a negative reference, as a foil. Elements relating to neo-ethnicity thus mix with the affirmation of a withdrawal from the mainstream of Northern societies. One of the difficulties with interpreting these dynamics is that the attitudes and behaviour of young generations of immigrants from the South do not completely lie in the domain of religion, neither in the sense of imposed attitudes and practices, nor in the political domain, nor also in respect for values and norms, but in the intersection between these levels. The other difficulty relates to the limits to information sources.
Study field, definitions and sources
5Some of the most eminent sociologists of religion, who, like Peter Berger in the 1960s, strongly emphasized the generality of the secularization process, have radically changed position, claiming a “re-enchantment” of the world.  It seems sensible to be more specific about the evidence and time horizons. Understanding contemporary trends firstly supposes looking beyond the fragmentation of religion. Measuring practices, albeit reported though statements, concerns a field that is more limited than that of beliefs. But in the Western world, some authors argue that dissociation between beliefs and affiliations and/or practices makes understanding religion as a whole problematic (Davie 2002; Hervieux-Léger 1999). To understand contemporary religiosity they find it necessary to broaden the definition of religion to embrace the world of beliefs independent of the idea of transcendence or, a fortiori, reference to a deity. While this universe of belief is a legitimate area of interest, there is no substitute for the study of religion as a set of obligatory practices, institutions and forms of devotion and reference to sacred texts. It is this religion, in the traditional, neo-Durkheimian sense to use Charles Taylor’s (2011) term, that I deal with here, deliberately setting aside broad spheres of belief.
6To account for the importance of “institutionalized” religion, three aspects are typically considered: religious affiliation, attitudes and practices. Firstly we can distinguish between those who claim to follow a religion as opposed to those who declare themselves agnostic or atheistic.  As for assessing the importance placed on religion or the religious education of their children by the former, here we are concerned with attitudes. We can also explore frequency of practice (taking part in worship, prayer, other acts of devotion, participation in charitable activities) among those who claim a religious identity (those who respond positively to the question: “Do you have a religion?”). Responding positively to the question “Do you have a religion?”, placing importance on this religion, respecting dietary restrictions or participating in worship are not governed by the same system of constraints. Declarations concerning practice, such as frequency of attending a service, are instructive as long as we remain within a fixed institutional framework. However, they create an illusory sense of accuracy if the context evolves and if we compare several countries, notably because the demands of these societies vary depending on their degree of religious pluralism.  These constraints do not impact on religious attitudes as much as they do on practices, and conversely declarations concerning attitudes are more marked by social desirability than practices. However, empirically, attitudes and practices form the same dimension and are extremely predictive of one another: depending on the degree of importance placed on religion—and without systematically accounting for the practices of the most fervent —, the probability of regular monthly practice changes completely. In Europe, even where one of the major aspects of secularization would be a comparative decoupling of attitudes and practices,  most studies show that the latter—attending worship, prayers, etc.—are highly correlated with subjective expressions of religiosity, such as the importance attached to religion and religious education (Halman and Draulans 2006). While the strong correlation between these items empirically justifies speaking of religiosity to designate all declarations concerning practices and attitudes expounding a positive stance towards religion, in the comparisons that follow we will however state whether the indicator used to represent religiosity stems from stated practices or the importance placed on religion (the subjective dimension of religiosity).
7In Europe, there are fairly abundant sources touching on the religious context and from which changes in practices and expressions of religiosity can be revealed. Religiosity is the subject of recurrent investigations from various international studies, three of which offer rich data: European Values Studies (EVS) 1981–2008/9, the European Social Survey (ESS) 2002–2008—which is a comparative survey protocol financed by the European Commission—and the International Social Survey Programme (ISSP) 1984–2009—archived at GESIS in Cologne.  Other than religious identity (faith) these large studies record standardized data on attending worship, prayers and past practices or those of parents and indices of the importance placed on religion, which allows us to paint a meaningful picture of recent developments. Data on the countries of origin of immigrants, and in particular on immigrants with African and Asian origins, are for the moment extremely sketchy. Certainly the World Values Surveys (WVS) published by the ASEP/JDS archives in Madrid furnished indicators for North African and Sub-Saharan African countries at varying dates during the 1981–2007 period, but these do not really allow us to consider the temporal dynamics for this region because in the region that interests us Turkey is the only country where there have been several surveys. 
8Regarding the practices and attitudes of immigrants and descendants of immigrants in France, conventional surveys have been based on relatively small sample sizes: they do not allow us to know the significance of religious affiliation and frequency of practice except in an approximate way. Claude Dargent (2010) has been able to overcome this obstacle through a combination of surveys from the Observatoire Interrégional du Politique (Interregional Policy Monitor). In addition, two large-scale surveys conducted in France at sixteen year intervals permit comparisons over time: the Mobilité Géographique et Insertion Sociale (MGIS) (Geographical Mobility and Social Integration) survey conducted in 1992, and the Trajectoires et Origines (TEO) (Trajectories and Origins) survey carried out at the end of 2008, by INED and INSEE. Religious identity and the level of practice and religious education of immigrants in France were thus observed in 1992 and 2008. The behaviour and attitudes of descendants of immigrants are sadly not known except for two ethno-cultural groups in the MGIS survey, and there was no measure that indicates religiosity in 1992. While we can compare the practices and descendants of certain groups between 1992 (MGIS) and 2008 (TEO), the variations between the question formulations relating to practices from one survey to the other makes this exercise problematic.  A comparison of the level of irreligion over time is possible in principle, because the formulations are identical, but in reality is difficult. Indeed, the percentages of “no religion” in 2008 seems to me to bear the hallmark of a “halo” effect due to the chain of questions that precede it: interviewees were asked if they placed importance on the religious education they received in their family, then: “Today, do you have a religion?” In MGIS, the first question is factual, interviewees were asked whether they “received a religious education when they were children,” then: “Do you have a religion?” There was no assessment of the importance accorded to the education received. Presumably those in 2008 who said they did not attach much importance to the education they received were more inclined to state that they had no “religion today.” Therefore, the noted increase in the percentage of irreligion in the space of sixteen years, which rose from 24% to 54% is perhaps exaggerated in view of the rather slow decline in religiosity that was experienced more generally over this period. 
9In 1992, MGIS formulated the question in an inclusive way: “Do you practise regularly, occasionally or never?”—and attempted not to favour Christians. Of these, indeed, the question must have been interpreted as regularly attending a religious service, which in France today means monthly attendance.  Among Muslims, it was the regular performance of prayers that seems to have been the general interpretation. We know that attending collective Friday prayers varies greatly by sex: men are required to do so, women are not. In Islam, visiting places of worship—which involves exposing oneself—is not obligatory for women, who must, in contrast, cover themselves in public to avoid being looked at.  Yet, the responses from men and women from the Maghreb and the Sahel to the question: “Do you practise …?” are of the same order of size in the 1992 MGIS survey (the interpretation of the question by interviewees does not therefore relate to attending worship). TEO, in 2008, asked “Do you attend a religious service?” offering a range of frequencies. There can therefore be no question of comparing absolute levels of practices from one faith to another, and the differences between the 1992 MGIS survey and the 2008 TEO survey must be carefully assessed. As for practices, respect of dietary restrictions was still addressed in the two surveys and can be compared. Wearing religious symbols was only mentioned in the TEO survey,  which was also the only one to pose the question on the importance placed on religion.
Religion of father by origin: Descendants of immigrants and the majority group in France (%)
Religion of father by origin: Descendants of immigrants and the majority group in France (%)Sample: Descendants of immigrants and the majority group in France, aged 18–50 years. The majority group is made up of those in France who are neither immigrants nor direct descendants of immigrants (having at least one immigrant parent).
Interpretation: Of the immigrants from the Maghreb, 85% state that their father is Muslim, 9% that they have no religion.
10For comparisons of attitudes and practices of immigrants and descendants of immigrants, three regions where Islam dominates were collated here—Maghreb, Sahel and Turkey—in comparison to two where it is Christianity that dominates— European countries, and the countries of central Africa and the Gulf of Guinea.  While there is a certain degree of pluralism in each cultural area, Table 1 shows that according to descendants of immigrants, a single faith dominates at home. It is only in France that an alternative option—declaring no religion—represents one-third of responses. It seems unlikely that immigration selects the most religious faction of the populations of the Sahel, Maghreb and Turkey where the dominant faith, at more than 80%, is Islam.
Importance of religion in the South, decline in practice in Europe
11A formal consideration of contexts implies having data on subjective religiosity in the south and east of the Mediterranean, which do not exist to my knowledge; some quantified indicators are nevertheless available from international surveys. To interpret immigrants’ levels of practice and attitudes, we must bear in mind the recent changes to religious contexts, both in France and overseas. Immigrants from Africa and Turkey have inherited a strong religious presence, both institutional and personal, from Africa and the Middle East. In the period beginning from the Arab defeats in 1967 and 1973, the weakening, if not the collapse, of secular ideologies (those of the Syrian and Iraqi Baath parties and the Algerian National Liberation Front) resulted in profound changes to attitudes and religiosity.  According to many observers a new relationship between religion and politics was established in Muslim countries, of which for two decades fundamentalist Islam has been and still is a strong expression (Mede 2002; Roy 2004; Keppel 2004; Redissi 2006). These changed were significant even beyond the epicentre of these upheavals. An accumulation of indicators—an increase in veil wearing and pilgrimages to Mecca, development of madrasas, growing references to Sharia law in civil codes, amendments to provisions promoting sexual equality in personal status codes—signified a revival of conservative religious influences in the south and east of the Mediterranean. Here, even though a convergence with the West was taking place— changes in fertility rates, increased school enrolment, development of economic modernization despite or because of oil revenues—religion reinforced its presence in civil life and sometimes within institutions. It manifests itself in a collective and diffuse way that statistical surveys on individual levels of practices do not account for well.
12As has been said, there are few statistical measures of religion in Africa. However, according to the WVS, the level of religious practice in the Sahel is very high: in Mali and Burkina-Faso in 2007 more than 75% of adults claim to attend a religious service once a month, and more than 95% in northern Nigeria in 1999. In central and Guinean Africa, monthly practice stood at 88% in Ghana in 2007 and more than 90% in southern Nigeria in 1999. At the turn of the of the 2000s, the proportion of those in North Africa, the Sahel and Turkey who attended a religious service monthly was 65% on average of 18–20-year-olds and 90% of those over 70.  Turkey bears witness to the situation of a Muslim country marked by secularization  from above and undergoing rapid modernization, where the high frequency of monthly attendance of worship has not declined in recent decades, although it varies by age.  We can thus, it seems to me, assume the perpetuation of a very high level of religious practice both in the countries of central and Guinean Africa and Sahelian Africa, North Africa and Turkey.  Whatever the reasons for, and diverse evidence of, this religious vitality in the South, this starting point cannot be ignored when considering the religiosity of African and Turkish immigrant minorities.
13This significance of religion contrasts with the largely secular context in Europe. From the point of view of religious practice, four large areas are distinguishable. The Protestant, Lutheran, Calvinist or Anglican countries (Denmark, Norway, Sweden and the United Kingdom) which can be grouped in terms of practice with the twin faith countries (Netherlands, Germany and Switzerland); the Catholic countries (Austria, Czech Republic, Slovakia, Slovenia, Hungary, Poland, France, Spain, Italy, Portugal, Ireland); the Orthodox countries (Serbia, Russia, Bulgaria, Romania, part of Bosnia, Greece, Cyprus); and the Muslim countries (half of Bosnia, Kosovo and Albania). From 1950 to 2010, there was an observable overall decline in stated practices and beliefs. Based on respondents’ statements about their parents, Franz Höllinger and Max Haller (2009) shows that monthly attendance of a religious service declined in Western Europe—Protestant and Catholic—between the 1930s–1950s and 1998. This decline was particularly pronounced in the Netherlands (–37 points) and in Austria, Spain and France (–25/–27 points). Comparative data compiled by Colin Crouch (1999) show that most practices—baptisms, church marriages and service attendance—declined in these European countries between 1970 and 1990. With the exception of the United Kingdom, where already low levels of practice rose slightly, religious service attendance  had declined systematically between 1970 and 1999. In France, according to Danièle Hervieu-Léger,  weekly attendance at mass among Catholics declined from 20% to 15% between 1960 and 1990. The decline in the belief in God among young people is very pronounced: Yves Lambert notes that the major caesura occurred around the 1970s, the percentage claiming they believed in God  declined from 75% on average for the years 1950–1970 to 45% for the years 1975–1990. The data collected by this author on practice in France between 1952 and 1992 show the decline in the proportion of young people attending worship on a weekly or monthly basis as clearly: practice fell away before the end of the post-war boom years (Trente Glorieuses),  in 1965 and in 1975–1977.
14The “snapshots” taken at the beginning of the 1980s or 1990s by the EVS indicate levels of religious service attendance that are higher overall than those recorded during the 2000s. These differences vary by age (Figure 1). In Catholic countries, where attending worship is on average higher than in Protestant countries, the decline took place during the 1980s. It is the 36–65 years age groups that contributed the most to declining rates: a drop of about 20 points throughout the period. In Protestant countries, from 1980 to 2000, the break in the slope that marks the threshold elevation of monthly worship attendance moves after 50 years, so that at the beginning of the 2000s, barely 10% of those under 50 attended a service monthly (Figure A1 in the appendix). The results from the ESS and ISSP surveys largely conform to these. During the 1990–2010 period, this decline in practice continued to take place both in Catholic Europe where the levels remained significant, at least according to the criterion of monthly church attendance, and in Protestant countries where it barely passed 20%. In countries of the Orthodox faith, which for a long time were in the Soviet orbit, another dynamic is observed. The levels of monthly practice run contrary to those in the West: their lowest levels were recorded in 1990 and they rose between 1990 and 2008 (see Figure A2 in the appendix). Though open to question because of the proximity in time between surveys, the observations provided by the WVS in 1997 and 2002 on Muslim Europe also suggest a rising trend in practice.
Worship attendance at least once a month by age group in European Catholic countries in 1981, 1990, 1999 and 2008 (%)
Worship attendance at least once a month by age group in European Catholic countries in 1981, 1990, 1999 and 2008 (%)
15While some observers of the religious phenomenon, in the narrow sense of established practices as adopted here, discern the seeds of desecularization, this has neither disrupted the level of practice in Europe—which has continued to decline except in Orthodox countries and the fringes of Muslim Europe—nor the frequency of claims of a religious identity. It seems to me that what has transpired, however, is a relative exhaustion of the cycle of secularization in France, as in most Protestant countries.
The religious vitality of people from Muslim countries
16Although it might be considered that the TEO survey overestimates irreligion, a comparison with MGIS suggests that the proportion of men who claim a religion has at least been sustained between 1992 and 2008 among immigrants from the Sahel and Maghreb, while it has declined for individuals from central and Guinean Africa, Turkey and southern Europe. It can also be noted that with the exception of immigrants from the Maghreb, religiosity has increased among women. Concerning rates of monthly practice, while there are marked differences between groups from Sahelian Africa and other Muslim countries, the contrast to the mainstream of French society is mostly greater. In 1992, the percentage of immigrants of both sexes who regularly practised was 31% of those from the Maghreb, 53% of those from the Sahel, 34% when they come from Turkey and 10% of the majority group incorporating people who are neither immigrants, descendants of immigrants or from the DOM (French overseas departments) (Table 2a).  In 2008, the percentage of those who attended a service monthly distinguished individuals from the Maghreb (14%), the Sahel (24%) and Turkey (22%) more clearly from members of the majority group: less than 5% of monthly participants (Table 2b).  If we see “regular practice” and “monthly attendance” at worship as equivalents, there would have been a decline in practice between 1992 and 2008 among immigrants from the Maghreb and Sahel, and stability among those from Turkey. The level of practice was however significantly higher among individuals of both sexes coming from central and Guinean Africa where Christianity largely dominates (more than 75% of descendants of immigrants say their parents are Christians).
Religious practices and religiosity: Immigrants and the majority group in France in 1992 and 2008 (%)
1992Sample: Immigrants from southern Europe, central and Guinean Africa, the Maghreb, Sahel, Turkey and the majority group in France, aged 20–50 years.
Interpretation: In 1992, 8% of men in the majority group in France say they regularly practise.
2008Sample: Immigrants from southern Europe, central and Guinean Africa, the Maghreb, Sahel, Turkey and the majority group in France, aged 20–50 years. For a comparison with MGIS, southern Europe is limited to Spain and Portugal.
Interpretation: In 2008, 4% of men in the majority group in France say they regularly practise.
Religious practices and religiosity: Immigrants and the majority group in France in 1992 and 2008 (%)
17In France, wearing conspicuous religious symbols, which has been the subject of several laws and decrees (in particular a law in 2004), has as a result become a political issue.  Female immigrants and descendant of immigrants from the Maghreb, Sahel and Turkey more often wear conspicuous symbols (the famous veil) than men. Among Europeans, whilst there are also more women than men who wear religious symbols, the difference is small.  No comparison over time on this point is possible. As for respect for dietary restrictions, there is an evident difference between Christianity and Islam, where prohibitions are a major part of practice. In additions, regarding dietary restrictions, a clear increase in following dietary prescription can be observed among immigrants of both sexes from the Maghreb, Sahel and Turkey between 1992 and 2008, while there is little change among immigrants from southern Europe or central and Guinean Africa.
18Between 1992 and 2008, irreligious attitudes have deepened among European immigrants and in the mainstream in France—more people say they have no religion and when they do have one it is rarely practised. Among immigrants from the Maghreb and Sahel, irreligiosity did not increase and while the number of those claiming to attend worship has declined, respect for dietary restrictions has tended to become widespread among immigrants from all Muslim countries. This continued high religious vitality is also found among immigrants from central and Guinean Africa both sexes of whom claim higher levels of practice, although the proportions of irreligious—admittedly very low in 1992—are on the rise.
19A comparison of 20–30-year-olds in the MGIS and TEO surveys (Figure 2) shows the changes concerning the same age groups over a sixteen-year interval. These two surveys provide indicators on the assertion of a religious identity and practice at the threshold of adulthood, in two distinct cohorts of immigrants on the one hand and descendants of immigrants on the other. However, the number of ethno-cultural groups is limited.
Sample: Immigrants and descendants of immigrants from Spain, Portugal, central and Guinean Africa, Algeria, Morocco, the Sahel and Turkey.
Interpretation: Algeria and Morocco have been differentiated to make a comparison between MGIS and TEO.
Descendants of immigrants
Descendants of immigrants
21For immigrants, the socialization context relates to the religious climate in their countries of origin at the point of their departure. Assessments by observers and in newspaper articles suggest a religious revival, or at least a strong religious vitality, as much among Muslims as among African Christians during the 1980–2000 period,  while religious practice continues to decline in the mainstream with the exception of Eastern Europe.  If the declarations of 20–30-year-olds are a reflection of socialization, this should result in an affirmation of religiosity among immigrants from African countries and a weakening of those from Europe, individuals from Turkey being in a mixed position because of the twin influences (Western European secular currents and the religious revival of the East).  We might have expected to find an increase in the secular influence among descendants of immigrants, but during the last two decades of the twentieth century they have been accompanied by an evolution in familial contexts towards more religiosity.
22Religious profession—the fact of claiming a religion—increased between 1992 and 2008 among immigrants from the Maghreb;  was stable among immigrants from Sahelian Africa, declined among immigrants from Turkey and Guinean Africa and collapsed among those from southern Europe (Figure 2a). Monthly practice increased as strongly among immigrants from central and Guinean Africa, remained stable among immigrants from the Maghreb, declined moderately among immigrants from southern Europe and Turkey and strongly among immigrants from the Sahel.  Among immigrants from Algeria this time, and although they are exposed to secular influences in France, at the turn of the century there was a revival in religious affirmation and stable rates of practice. As for descendants of immigrants from southern Europe, there was a significant decline in religious affirmation and, to a lesser extent, in practice (Figure 2b).
Profiles of practice and the importance placed on religion by age group
23While a comparison of absolute levels of practice of immigrants is difficult given the heterogeneity of sources,  that of age profiles is interesting for several ethno-cultural groups. We can outline changes in behaviour  in the recent period concerning monthly attendance of a religious service (Figure 3). In 1992, the latter increased among the youngest age groups as well as in the oldest; moreover it did so in a more accentuated manner among immigrants of Moroccan origin than among Algerian-born immigrants. Regarding those from the Sahel, a significant increase in practice among the 21–25 years age group in comparison to the 36–40 years age group can be observed (15 percentage points). Compliance with periods of fasting and dietary restrictions increases between 26–35-year-olds and the 46–50-year-olds among immigrants of Algerian, Moroccan and Turkish origin. In 1992 there was still no systematic indicators of practice of descendants of immigrants. 
Frequency of regular practice by age: Immigrants from the Maghreb and Sahel and the majority group in France in 1992 (%)
Frequency of regular practice by age: Immigrants from the Maghreb and Sahel and the majority group in France in 1992 (%)Sample: Immigrants from the Maghreb and Sahel, majority group in France, aged 21–60 years.
24Sixteen years later, the dynamics had changed in the TEO survey. In 2008, the frequency of monthly worship by the majority group in France was still lower for young people—it is around zero for the youngest—than among those aged 46–50 years,  where, however, it is less than 10% (Figure 4). But among descendants of immigrants from the Sahel, Maghreb and Turkey, it rises in contrast when looking at the youngest age groups. 
Attending worship at least once a month: Immigrants, descendants of immigrants from the Maghreb, Sahel and Turkey, and the majority group in France in 2008 (%)
Attending worship at least once a month: Immigrants, descendants of immigrants from the Maghreb, Sahel and Turkey, and the majority group in France in 2008 (%)Sample: Immigrants and descendants of immigrants from the Maghreb, Sahel and Turkey, majority group in France, aged 18–50 years.
25As regards dietary restrictions, while a strict comparison between 1992 and 2008 is not possible, the questions are similar enough for changes depending on age to be apparent. In 2008, respect of dietary restrictions acts as a strong norm for immigrants. Among descendants of immigrants, this norm is further radicalized in the young ages groups: the levels of respect for restrictions ranges from 60% to 90% among the oldest, reaching consistently above 90% among the youngest. There was no question on the wearing of conspicuous religious symbols in the 1992 survey, which is probably symptomatic, and neither can we compare the profile of this practice from one group to another in 2008. The frequency of wearing religious symbols by individuals who had migrated as adults is practically unchanged depending on age group, with the exception of a slight increase for the youngest Turkish immigrants. But among the youngest descendants of immigrants from the Maghreb and Sahel, who have thus been socialized in France, the wearing of religious symbols rises by 20 points.
26Let us consider, to complete this review, an indicator of attitudes depending on age in 2008, defined by the question “How much importance do you place on religion in your life today?”.  In the reference population, as we move towards the oldest age groups, the importance placed on religion increases—albeit slightly as we had to omit those aged over 50 years for reasons of comparability.  In contrast, among the populations from Algeria, Morocco and the Sahel, the importance placed on religion is greater among young people (Figure 5). It is as if the factors encouraging secularization in the majority group do not affect young people from immigration from the South but socialized in France. Most groups of young immigrants from Muslim countries also buck the trend in declining religious attitudes. The importance placed on religion, by nature not subjected to prescriptions, enables an easier comparison between sexes and faiths than indicators of practices. So I will now focus on this synthetic indicator: although taking liberties with the terminology, I will use the word “religiosity” to designate it.
Importance placed on religion: Immigrants arriving when under 17 years, descendants of immigrants and the majority group in France in 2008 (%)
Importance placed on religion: Immigrants arriving when under 17 years, descendants of immigrants and the majority group in France in 2008 (%)Sample: Immigrants arriving in France aged under 17 years, descendants of immigrants and the majority group, aged 18–50 years. For the Sahel, we have grouped together those aged 36-40 years and 41-45 years.
Interpretation: The points represent groups that number more than 30 individuals.
27All things being equal, while women place greater importance on religion than men, this importance is independent of level of education in the six ethno-cultural groups under consideration (Table 4). The odds ratios associated with the importance placed on religion are six times greater among 18–20 year old immigrants and descendants of immigrants from the Maghreb or Sahel (four times for people from Turkey), in comparison to the 46–50 year old cohort; while of the people from Europe or central Africa/Gulf of Guinea, the age effect is not significant (there is however a significant reduction in religiosity among descendants of immigrants between the 20–30 years and 46–50 years age groups) (Table 4).
Importance attached to religion depending on origin and various socio-demographic characteristics: Immigrants, descendants of immigrants and the majority group in France in 2008 (%)
Importance attached to religion depending on origin and various socio-demographic characteristics: Immigrants, descendants of immigrants and the majority group in France in 2008 (%)Sample: Immigrants and descendants of immigrants from Europe, central and Guinean Africa, Maghreb, Sahel and Turkey, the majority group in France, aged 18–50 years.
Interpretation: 9% of men and 15% of women in the majority group place importance on religion in 2008.
Odds ratio of logistic regressions concerning the importance of religion: Immigrants, descendants of immigrants and the majority group in France in 2008
Odds ratio of logistic regressions concerning the importance of religion: Immigrants, descendants of immigrants and the majority group in France in 2008Sample: Immigrants and descendants of immigrants from Europe, central Africa and the Gulf of Guinea, from the Maghreb, Sahel and Turkey and the majority group in France, aged 18–50 years.
Notes: *** p < 0.001; ** p < 0.01; * p < 0.05. # As an alternative we have estimated a model where the cut off for age at arrival is 18 rather than 16: the results are analogous.
Interpretation: The values indicated in the column are odds ratios (exp(b)) of coefficients: above 1, indicates a positive effect on the “importance placed on religion,” less than 1 indicates a negative effect.
28In the most traditional sociological perspectives, religious attitudes are based on the idea that a person possessed of little knowledge and few skills attributes his life’s good fortune and misfortune to supernatural forces and/or a divine will. Conversely, a high level of knowledge is cited as the source of a “disenchanted” vision of the world, which can weaken religious beliefs and erode practices (Weber  1993; Housman and Mascini 2002). An increase in the level of education is also linked with a rationalization of the perception of the world, to the rejection of the idea of a universe created by a supposedly single will. Thus, in France, public education was particularly implicated in the decline in religious attitudes and contributed to discrediting the creationist idea. Yet, surprisingly, the effect of the education qualification variable on the subjective religiosity of descendants of immigrants is very weak.  Certainly, in a few groups, the most well qualified immigrants and descendants of immigrants tend to be less religious than those with no qualifications, but this is not confirmed among the groups of Maghrebis or those of central African origin (nor in the majority group). The better education of young people should accentuate the age effect: it does not.
29What might the factors be that could explain this apparent reversal of the trend among immigrants and descendants of immigrants from the Maghreb, Sahel and Turkey? The goal of this analysis is to understand what could determine the importance placed on religion among the young and masculine segments of these groups. Concerning immigrants, one of the interpretations of the decline in subjective religiosity with age sometimes advanced is the length of exposure to a secularized society (France in this case). Might this interpretation also be true for the descendants who have always been exposed to the influence of the host society? Why is the revival a masculine one and one that is specific to young people whose parents come from Muslim countries?
30To account for what can be observed, one could broadly assume that religious attitudes are formed both by socialization within primary groups—e.g., families and neighbourhood friends—and in response to current needs—e.g., the uncertainties of life, fear of death, need for recognition. Without being able to embrace the full spectrum of determinants—which relate to received influences and are formed in response to current needs—it is generally assumed that what is transmitted is diluted as it is passed down from immigrants to descendants, from parents to children. Attitudes that form in response to perceived needs can increase with age, or, when they reflect social anxiety about the future, be more pronounced among young people.
31To understand the importance placed on religion, we will consider factors relating to three areas—health, occupational situation and insecurity on the one hand and determinants related to inheritance and familial socialization on the other, and finally the historico-religious context of transitions into adulthood. These areas are less alternative explanatory factors than they are poles of attraction for the collection of explanatory factors within which the contextual influences of societies of origin and host countries are differently weighted. That which relates to living conditions in France—in particular poverty and ethnic segregation—is thus possibly related to three areas.
A response to adversity: poverty, insecurity and ill health
32The importance placed on religion is great in countries where income inequality is relatively high.  The idea that it is underpinned by anxiety resulting from economic insecurity and weakened material security, such that life’s difficulties are perceived as “fate,” is fairly common. By analogy, one might believe that being in an insecure occupation, having experienced lengthy periods of unemployment or being in ill health, induce feelings of vulnerability that can accentuate religious needs. Conversely, a relatively prosperous material life provides a feeling of mastery over one’s existence, altering religious needs. Independent of direct experience, insecurity created by high local unemployment and by noticeably deteriorating surroundings and declining guarantees associated with a wage society can also accentuate the need for metasocial guarantors and stimulate a revival in subjective religiosity.
33As evidence of personal economic insecurity I used a variable that tracks the employment situation of interviewees over the past five years, distinguishing those who were employed from those continuously self-employed, those in training and those who had been more or less continuously unemployed (short-term employment, long periods spent looking for work), those who have experienced more or less lengthy periods of economic inactivity (or all other discontinuities in employment) and, finally, housewives. I also used data on the level of income per unit of household consumption. Concerning the insecurity of living environments, I used the residential IRIS  ranking with respect to unemployment. I used a measure of the concentration of the immigrant population in residential IRISes, which, while it is mostly an indication of ethnocultural segregation, can also indirectly relate to insecurity. Finally, to give concrete expression to feelings of insecurity, I took into account the interviewees’ assessments of their own health.
34There is very limited empirical support to be found in the TEO for the hypothesis that, on an individual level, religious attitudes are nourished by insecurity and unemployment.  An average state of health is moderately associated with greater religiosity among immigrants from central Africa and the Gulf of Guinea and among Europeans, but it is not among immigrants from Muslim countries. As for occupational experience in the past five years, European immigrants who have had several employers devote more importance to religion than even those who had experienced phases of looking for work. Among Guinean and central African immigrants, having several employers and phases of unemployment is associated with devoting greater importance to religion—this is in line with the expected direction. Of those originating from the Maghreb, the Sahel and Turkey, experiencing employment insecurity has no impact; it is being a housewife that is, however, associated very clearly with greater religiosity. Subjective religiosity also varies with income per household consumption unit: it is greater among members of poor households, in particular Sahelian, Maghrebi and Turkish ones. In sum, individual characteristics of social insecurity and physical frailty are not associated with greater religiosity.
35In contrast, collective characteristics are associated. Thus “zones urbaines sensibles” (ZUS),  which are both poor and characterized by strong ethnic segregation and have characteristics that relate to insecurity and the isolation of ethno-cultural communities. The level of religiosity of immigrants living in neighbourhoods characterised by unemployment is greater than those in households living elsewhere. This suggests that there are mechanisms at work other than individual vulnerability. For immigrants from a minority faith, an environment where there is a strong immigrant presence, where the proportion of couples with the same origins is high and where sociability is local,  creates religious synergies. Conversely, in mixed neighbourhoods, where there are more mixed couples, where children are in a situation of normative pluralism, the importance placed on religion tends to decrease.
36While the influence of neighbourhood on religiosity is weak or null for descendants of immigrants from central and Guinean Africa and European immigrants’ descendants (Figure 6),  immigrants and more especially descendants of immigrants from the Maghreb, the Sahel and Turkey living in segregated districts place more importance on religion than those dispersed in districts with little segregation. The type of district interacts with the composition of parental household on the religiosity of those from the Maghreb, Sahel and Turkey but not on others. Thus, the variations in subjective religiosity associated with cultural isolation combine with those associated with primary socialization contexts. This leads us to question the value systems of young people.
Importance placed on religion depending on whether a person has one or two immigrant parents, by neighbourhood of residence and cultural origins: descendants of immigrants in 2008 (%)
Importance placed on religion depending on whether a person has one or two immigrant parents, by neighbourhood of residence and cultural origins: descendants of immigrants in 2008 (%)Sample: Descendants of immigrants from Europe, central and Guinean Africa, the Maghreb, the Sahel and Turkey, aged 18–50 years.
Interpretation: 19% of descendants of immigrants of whom one parent is of European origin, living in a non-ZUS neighbourhood, state that religion is “important” or “very important.”
Socialization, normative dissonance and time of arrival in France
37We know that religious dispositions acquired during childhood, or at least imprinted during the prime of youth, have a certain durability (Vertovec and Rogers 1998): this period is thus of particular interest. Lots of work on immigration emphasizes the importance of heritage internalized at the point of reaching adulthood (Foner and Alba 2008). At young ages, socialization is above all familial, during adolescence familial and extrafamilial influences combine. Immigrants have been exposed to a variety of norms: those received as children and adolescents in their countries, with various contributions form the host country have increasingly intermingled.  For immigrants who arrived when young, as for descendants, early extrafamilial influences relate to two poles, the country of origin’s as well as the host country’s culture, which can be congruent or dissonant both in terms of intensity of practice and faith orientation. The hypothesis of a “passive” interaction would lead us to believe that the individuals in religious dissonance would be inclined towards a more temperate religious attitude the longer they have been exposed to the secular attitudes of the host country.
38Similarly, the degree of subjective religiosity should logically be weakened by potential dissonance between familial sources of influence. Thus, religiosity is more marked if both parents are immigrant and not just one of them. The congruence in parental beliefs, quite rare when one parent is Christian or irreligious (10%–20%), becomes very rare when one parent is Muslim (2%–5%). In all cases, this faith dissonance induces an attenuation of the importance placed on religion, which varies depending on region of origin: its impact is greater on groups originating from Muslim countries, while its impact on groups from Europe or central or Guinean Africa is weaker or nonexistent.
39Furthermore, variations in subjective religiosity among immigrants and descendants of immigrants follow three distinct patterns depending on origins (Figure 7). Among descendants of European immigrants, religiosity is twice as important for those whose father or mother is Christian rather than irreligious or of a different faith: “Christians” are only produced in families where one of the parents declares themselves Christian, but the congruence of the mother’s or father’s faith has no impact. Among descendants of immigrants from central or Guinean Africa, not only does the congruence of messages provided by parents have little importance but the fact the father of the respondent calls himself Christian or otherwise has little effect on the interviewee’s religiosity, as if the importance of religion depended little on individual socialization conditions.
Importance of religion by region of origin, religion of father and the congruence of parents’ religions in 2008 (%)
Importance of religion by region of origin, religion of father and the congruence of parents’ religions in 2008 (%)Sample: Descendants of immigrants from Europe, central and Guinean Africa, the Maghreb, Sahel and Turkey, aged 18–50 years.
Interpretation: 26% of descendants of immigrants originating from Europe with a Christian father but a non-Christian mother, claim that religion is “important” or “very important.”
40Among descendants originating from the Maghreb, Sahel and Turkey, religiosity is much more pronounced when fathers are Muslim (in consonance with the dominant faith) than when they are not. Moreover, congruence between the religious choices of mothers and fathers has a strong impact on the religiosity of the self. Familial socialization thus has very different effects from one cultural space to another depending on faith and the congruence of parental belief.
41Most often, socialization is conceived of as “passive”: it can be considered actively, and religious attitudes, particularly those of second generations, can be considered to be self-constructions in response to life situations formed by received legacies. Unable to distinguish between age, period and cohort effects with a single survey, to understand the interactions between received and adopted values I distinguished sub-groups depending on the moral context during adolescence and explicitly introduced interactions with age and context characteristics.
42In fact, the models in Table 4 treat all individuals with the same geocultural origins uniformly, excluding interactions between periods of arrival, generation and type of ancestry (mixed or not) and those between these characteristics and age. Yet it should be noted that the models used do not all have the same “explanatory strength” from one ethno-cultural group to another: they are weak for immigrants from Europe (models 1 and 2), intermediate for immigrants from countries in Christian Africa (model 4) and significantly stronger for immigrants from Muslim countries (models 3 and 5). The comparison of the coefficients of the variables used shows that those that contribute the most to differentiating this explanatory power are age, status with respect to migration and ancestry.
43The formalization adopted in Table 5 enables the distinction of different sub-groups within each ethno-cultural group by distinguishing socialization contexts and the testing of their potential interaction with age. These sub-groups were constructed by intersecting three variables—the relationship to migration (i.e., first or second generation), the identity or difference in origin of two parents and the period of arrival.  For immigrants, the significance of period of arrival defines different experiences: it separated those who arrived before 1974 (the end of labour immigration), between 1974 and 1983 (before the implementation of the long-term residence permit in 1984, which reduced the risks of departures and returns), and the 1984-1992 period and that after 1992 (the moment of radicalization of tensions between Muslims and non-Muslims linked to the first Gulf war).  For descendants of immigrants, their parents’ arrival date has no direct meaning. However, it can provide an indicator of the climate of religiosity in the family, depending on whether parents arrived in 1962 or earlier, between 1963 and 1974, from 1974 to 1983 or after 1983. Each arrival cohort contains a diversity of ages, which enables the estimation of specific “age/period of socialization” effects.  The random part of the models is thus based on 15 sub-groups.
44The “age” variable is treated as continuous, but while age is often interpreted in terms of life-cycle position, here a different perspective is adopted. Thus, for example, being 20 years old in 2008 (the date of the survey) means having been an adolescent at the beginning of the second Gulf war and from this perspective that denotes a series of moments defined by their moral and political tone rather than aging.  To account for potential interactions between this variable and the sub-groups’ characteristics, I used two types of modelling (Appendix Table A3). In the first, models 1 and 3, the “period of socialization” effect is identical in each subgroup and thus independent of the immigrant/descendant of immigrant distinction, type of ancestry (one or two immigrant parents) and period of arrival; in the second, models 2 and 4, the effects of the “age/period of arrival” variable on religiosity are allowed to vary from one subgroup to another within the same region of origin.  Table 5 presents the deviations from the mean of the importance placed on religion and its profile by “age/period of socialization” for the two origin sub-groups according to models 2 and 4.
Deviation from the mean of the logit of the importance placed on religion (u_o) and the slope depending on age, whether immigrant or descendant of immigrants (see Appendix Table A4)
Deviation from the mean of the logit of the importance placed on religion (u_o) and the slope depending on age, whether immigrant or descendant of immigrants (see Appendix Table A4)Sample: Immigrants and descendants of immigrants from Europe, the Maghreb, the Sahel and Turkey, aged 18–50 years. The control variables not presented in this table are sex and age of respondent, father’s faith, religious congruence between parents, mother’s activity, household income, the number of immigrants in the neighbourhood.
Interpretation: In comparison to the average level of stated religiosity, the youngest descendants of immigrants from the Maghreb, Sahel and Turkey from parental cohorts having arrived before 1962 or between 1963 and 1973, place distinctly more importance on religion than the average sub-groups (+1.72 and +1.97 respectively) when they have two immigrant parents.
45While the level of religious practice has declined in most countries of origin of European immigrants, it remains higher than in the majority group in France, above all in the old cohorts. According to model 2, this does not however systematically produce positive deviations from the mean among European immigrants and their cohorts (Table 2, model 2, deviations from the mean).  Of the descendants with a single immigrant parent from a European country, the deviations from the mean are evidence of a decline in religiosity, being systematically negative, with relatively subtle “age/period of socialization” effects (the coefficients rarely exceed |0.02|). The effect of parents’ period of arrival on religiosity becomes positive when both parents arrived after 1962, and it is also not very age sensitive. In sum, the differences in religiosity as a function of “age/period of socialization” are low for descendants of European immigrants.
46For immigrants from the Maghreb, the Sahel and Turkey, analysis suggests that the practices and attitudes of those who arrived in the most recent waves reflect the growing assertion of Islam in societies from the South (Table 5, model 4: the deviations from the mean increase with the arrival period).  However, for descendants of immigrants from the Maghreb, the Sahel and Turkey, the average levels of religiosity defined by the cohort of their parents’ arrival do not characterize attitudes well. The religiosity of descendants of two immigrant parents from Muslim countries is extremely heterogeneous depending on period of socialization and parents’ arrival cohort: the contrast between younger and older being even more accentuated in cohorts of parents who arrived between 1963 and 1973 (large deviations from the mean and steeply negative slopes implying a significant decline in religiosity of the eldest: Table 5, model 4, italics in the final column). These contrasting dynamics are illustrated in Figure 8: with respect to the group average, the estimated level of religiosity of descendants of a single immigrant parent from the Maghreb, Sahel or Turkey is relatively low and increases with the length of socialization. It is significantly higher than the average for those who have two immigrant parents (with particularly large differences depending on period of socialization when their parents arrived before 1974). Living in a ZUS only induces differences in levels.
Deviations from the mean of the logit of the probability of devoting importance to religion (random part), depending on ancestry, period of arrival of parents and neighbourhood (model 4b, appendix Table A4): Second generation originating from the Maghreb, Sahel and Turkey
Deviations from the mean of the logit of the probability of devoting importance to religion (random part), depending on ancestry, period of arrival of parents and neighbourhood (model 4b, appendix Table A4): Second generation originating from the Maghreb, Sahel and TurkeySample: Descendants of immigrants from the Maghreb, Sahel and Turkey, aged 18–50 years.
Note: For illustration, profiles by “age/period of socialization” of the logit of religiosity for certain sub-groups are represented (model 4b, appendix, Table A4).
Interpretation: Profiles by “period of socialization” of the logit of religiosity for certain sub-groups are represented (model 4). The less steep slope for the cohort of those with two immigrant parents who arrived between 1974 and 1983 indicates a lessening of the deviation in attitudes between older and younger siblings who grew up in a context of rising Muslim religiosity, in comparison to descendants from older cohorts.
47Parents arriving before 1974 who left their countries before the Islamic revival, exhibit a certain religious detachment through the values they promote in their children. One might think that children of immigrants aged around 50 in 2008 entered adult life in a family atmosphere of relative irreligiosity: it is this distance from religion that is recorded among the eldest in the second generation. The parents’ arrival date only however defines the parameters of the context of socialization—the familial religious tonality—whose effect is supplanted for the youngest by extrafamilial influences determined by the moral context of the host country during adolescence.
48The objection that could be made to this interpretation of increased religiosity is that the secular influence has not had enough time to affect these young people. But then this secular influence should also operate on Christians of African origin and this is not the case (see appendix, Figure A5. TEO. Importance placed on religion: persons originating from central and Guinean Africa). It is young Muslims who devote a lot of importance to religion and not young people in general, and it is hard to reduce this religiosity to a heritage transferred from elders, who, on the contrary, when they have been in France for a long time, devote less importance to religion. In the recent period when the affirmation of Islam has been strong and at the same time hostility to Muslims has been marked, the weakness of transmitted religious legacies (less importance of religion in cohorts of parents who arrived before 1974) does not preclude significant religiosity among the youngest descendants of these cohorts, which is constructed based on age group influences, neighbourhoods frequented (ZUS) and the global context. This leads us to consider a process where the importance placed on religion, rather than weakening in older age groups, is reinforced among young descendants from Muslim immigration, socialized in France, having two immigrant parents and living in segregated districts.  They are distinguished from their parents by this religious affirmation,  by which they often adopt a contrary position by conspicuously adopting Islamic postures if not practice. It is as if the second-generation young people have received and internalized the religious revival in their country of origin more than their parents who left less fervent contexts. Hostility to Islam from mainstream France thus produces an intergenerational and/or transnational proximity that runs counter to the stance noted by Abdelmalek Sayad ( 2006) nearly thee decades ago. From this stems the idea of interpreting their strong religiosity as identity construction in reference to their situation in France and to the evolution of Islam outside of France.
Religiosity and identity assertion
49To confirm or deny the existence of a link between the importance placed on religion and identity construction among immigrants and descendants of immigrants, variables that could account for this identity assertion have been taken into account in a final series of analyses (Table 6).  Identification with a country of origin or the host country was measured in the TEO survey by agreement or disagreement with the assertion “I feel French,” “I feel Turkish,” etc,. which reflects membership of each of the national groups. I used a variable that combines French and gentilic (identity associated with parents’ birth country) identities, so as not to prejudge exclusive identities: one can have a feeling of being both from “here” and from “there.” I also took into account the degree of transnational openness of immigrants and descendants of immigrants. The notion of transnationalism refers to the large variety of social, economic and political practices and orientations exercised beyond the borders of the country of residence. The most economically active individuals also often have strong transnational orientations in symbolic, social and political domains. In order to account for the intensity of practices two scores were calculated: one encompassed a practical dimension (journeys to countries of origin, sending money, etc.) the other covered a symbolic dimension (exchanges with family in the home country, attention to what is going on there, desire to be buried there). A high transnationalism score has little effect on the frequency of the claim: “I feel French” while it greatly increases gentilic claims corresponding to the origins of parents. The two variables are not however substitutable, which is why I kept both of them together (the components of transnationalism measured are described in the appendix).
Importance placed on religion and gentilic identity in France: Odds ratios of variables concerning feelings of having been discriminated against, identification and transnationalism, descendants of immigrants
Importance placed on religion and gentilic identity in France: Odds ratios of variables concerning feelings of having been discriminated against, identification and transnationalism, descendants of immigrantsSample: Descendants of immigrants from Europe, central and Guinean Africa, the Maghreb, Sahel and Turkey, aged 18–50 years. The following variables were controlled for: sex, parents’ arrival period, age, income, mother’s activity, father’s faith, congruence of parents’ faith, segregation of residential neighbourhood.
Interpretation: Odds ratio of descendants of immigrants from the Maghreb, Sahel and Turkey who do not feel particularly French (“French + or -) but who do not identify with their parents’ country of origin is 2.53 in reference to those who claim to feel “French ++” and feel distanced from their parents’ country of origin.
Note: *** p < 0.001; ** p < 0.01; * p < 0.05.
50The feeling of having been discriminated against plays a minor role in identity construction: the upsurge of religiosity seems to have occurred during recent decades among members of minorities who, without feeling general hostility,  need to assert their otherness. Elements of assertions of this cultural background also stem from the appeal of the media and from interpersonal exchanges via the Internet. The influence of “over there” is not only borne by parental heritage: in the current contexts it takes on a character that is both local and global, overriding national contexts. Thus, young people from Maghrebi and Sahelian immigration that watch Al-Jazeera, African television channels or Maghrebi internet sites (behaviour captured by the symbolic transnationalism score) are influenced by visions of the world spread from the Ummah and appear the most religious. The greater the native influences the greater the asserted religiosity. However, it is not a given that this should be seen as just a passive expression of the socialization of the subject who simply has a role as a filter of received influences. Analysis invites us to depart from this passive determinism and consider religiosity not just as something impressed on people by socialization, but also as a response and affirmation.
51* * *
52Among individuals originating from African countries, and in particular those that are majority Muslim, an exaltation of religiosity can be observed, at least in the form of claims of a religious affiliation, even if this does not always greatly translate into practice. Young descendants of two parents who immigrated from the South develop religiously in ways contrary to the mainstream of the host society, sometimes distancing themselves from parental influences. In France, their religiosity is amplified rather than attenuated. This amplification can be interpreted as a form of confrontational or minority socialization redolent of Alejandro Portes and Ruben Rumbaut’ selective assimilation (Portes and Rumbaut  1996, 2001), implying, at least in certain areas, maintaining a distance from the host country even if it is not an oppositional culture. Work carried out in recent years shows that it is not an “either or” situation but that there are configurations where native identities promote the material and symbolic acculturation of immigrants: this is the integrative aspect of minority identity assertion. 
53At the turn of the twenty-first century, the effect of secularization on young age groups of descendants of immigrants from Muslim countries has been supplanted by an assertion of identity and/or adherence to the Muslim Ummah symbolized also by transnational practices. The attitudes and behaviour of these young people from immigration from the South depict a space where religious affirmation is chosen as a mode of identity, mobilizing a heritage and a position within a context saturated by tension due to the emergence of political Islam. Their religious revivalism should be thought of as an aspect of the renewal of identity assertion in a world experiencing a rapid dramatic change in North–South relations and a Europe coming to the end of its cycle of secularization.
54Today, we can no longer assume that descendants of immigrants want to resemble “us” in every aspect, nor that differences in identity are simply evidence of incomplete integration. One can hope to succeed in the countries of the North— through good education, mastering the vernacular language and understanding how the administration functions —while having a distinctive identity. In becoming a reality, the cultural diversity of European societies has tended to discredit a vision of integration as erasing differences in favour of a perspective where ethnic or cultural identities tend to recompose themselves by selective borrowings and a plurality of spaces and references. History teaches us the proscription of beliefs induces the maintenance and sometimes reinforcement of specific identities among those it is directed at.  In Europe today, the creation of minority identities has become a major issue of what might be called, in line with Jürgen Habermas (2000), post-national societies: xenophobic tensions illustrate the difficult emergence of normatively open social groups. Even when one is aware of the strength of these tensions, interpreting cultural—linguistic, religious and moral—differences does not mean measuring the distance from a central core that is the reference, but means understanding the normative recompositions without thinking of them as residues. To succeed here and to lose part of one’s identity, or reinforce the otherness of one’s identity at the cost of sufficient happiness in the host country, are not the only alternatives.
1 – Definitions
55The responses concerning the importance placed on religion and on religious education reflect what William James ( 1985) calls propositional attitudes. These are statements that tell of the degree of adherence to or rejection of propositions on the world or the relations between the subject and the world without explicitly denoting practices or experience. They do not have the same status as responses describing practices. 
56Questions used in the surveys on immigrants and descendants of immigrants in France to describe the religiosity and religious behaviour.
P19—Do you have a religion? (no indication which)
P191—Do you practise regularly, occasionally, never?
P20—Do you respect fast periods or dietary restrictions whether out of respect to religion or cultural traditions? (1 yes, 2 no, 0 no restrictions)
P18—Did you receive a religious education when you were a child? Yes/no
R_relsoi—Do you currently have a religion? (+ which)
R_culte—Do you take part in religious ceremonies? (1p/w, 2, 1 or 2p/m…)
R_impvie—How much importance do you currently place on religion in your life?
R_miam—Do you respect your religion’s dietary restrictions in your daily life? (1 always, 2 sometimes, 3 never, 4 there are none)
R_ostent—In your daily life, do you wear clothing or jewellery that symbolizes your religion?
R_impedu—How important was religion in the education you received? (1 not at all, 2 not very, 3 fairly, 4 very)
58The correlation between these items that describe religious attitudes and practices is given by the Crombach alpha value (setting aside the R_relsoi question).
59Average covariance between the items: 0.36; number of items: 5; coefficient: 0.62.
2 – Additional data
EVS. Attend worship at least once a month, by age, Protestant countries in Europe (%)
EVS. Attend worship at least once a month, by age, Protestant countries in Europe (%)
EVS. Attend worship at least once a month, by age, France (%)
EVS. Attend worship at least once a month, by age, France (%)
TEO. Wearing of religious symbols by composition of neighbourhood of residence and origins, women (%)
TEO. Wearing of religious symbols by composition of neighbourhood of residence and origins, women (%)Sample: Women immigrants and descendants of immigrants, majority group, 18–50 years.
Note: The categorization of neighbourhoods on the x-axis is the result of statements from respondents and is strongly correlated with the percentages of foreigners measured in the census.
TEO. Importance placed on religion. People from central and Guinean Africa (%)
TEO. Importance placed on religion. People from central and Guinean Africa (%)Sample: 18–50 years.
TEO. Importance placed on religion by ethno-cultural group: Immigrants, descendants of immigrants and the majority group in France in 2008
TEO. Importance placed on religion by ethno-cultural group: Immigrants, descendants of immigrants and the majority group in France in 2008Notes: *** p < 0.001; ** p < 0.01; * p < 0.05. The following variables are controlled in these regressions: health, occupational insecurity, household incomes, and neighbourhood context.
TEO. Importance placed on religion by country of origin: odds ratios ages and arrival period: immigrants and descendants of immigrants
TEO. Importance placed on religion by country of origin: odds ratios ages and arrival period: immigrants and descendants of immigrantsSample: Immigrants and descendants of immigrants, aged 18–50 years. Variables controlled: income, mother’s activity, father’s faith, congruence of parent’s faiths and segregation of neighbourhood.
Interpretation: The odds ratio of female immigrants from Europe is 1.38, in reference to men of the same origin.
Note: *** p < 0.001; ** p < 0.01; * p < 0.05.
TEO. Importance placed on religion: regression coefficients for the mixed logistic models: immigrants and descendants of immigrants from Europe and Maghreb/Sahel/Turkey
TEO. Importance placed on religion: regression coefficients for the mixed logistic models: immigrants and descendants of immigrants from Europe and Maghreb/Sahel/TurkeySample: Immigrants and descendents of immigrants, 18–50 years.
Note: *** p < 0.001; ** p < 0.01; * p < 0.05; # p < 0.1.
Interpretation: The fixed section presents coefficients associated with a direct influence from interpretive variables on the variable of interest (importance placed on religion). The variable section represents the standard deviation of the level of religiosity between sub-groups defined by generation, ancestry, neighbourhood, etc., and the standard deviation of interactions between age/period of socialization and sub-groups (these interactions are interpreted as “slopes”).
TEO. Deviation from the mean of the logit of the importance placed on religion (u_o) and slope coefficients by age (random part, model 4a) Immigrants from Maghreb/Sahel/Turkey
TEO. Deviation from the mean of the logit of the importance placed on religion (u_o) and slope coefficients by age (random part, model 4a) Immigrants from Maghreb/Sahel/TurkeySample: 18–50 years.
Interpretation: see Figure 8.
3 – Transnational dimensions in TEO 2008
601) Economic dimension: Ownership of property or land outside of metropolitan France, investment in a trade or company outside of metropolitan France, regular financial support of a household outside of metropolitan France, financial contribution to a collective project outside of metropolitan France (school, clinic, etc.).
612) Relational dimension: Personal contacts by letter, telephone or the internet outside of metropolitan France, visits to the respondent’s or parents’ home region, use of media from the respondent’s or parents’ home region, interest in the political life of the respondent’s or parents’ home region.
623) Symbolic dimension: Desire to be buried outside of metropolitan France, wish to leave to live outside of metropolitan France, feeling of belonging to region of origin or that of parents.
There is no consensus on the use of the words “integration” and “assimilation,” still employed in the literature (Rumbaut 1996; Tribalat and Kaltenbach 2002).
In an interview for Christianity Century (1997), Peter Berger declared: “I think what I and most other sociologists of religion wrote in the 1960s about secularization was a mistake …”. Without explaining what the mistake was, he suggested we should not be surprised: “the interesting question is not how do you explain fundamentalism in Iran? But, why is Western Europe different?”
The distinction between those who question the idea of transcendence (agnostics) and those who question its personification as a deity (atheists), which for thought is of great importance, is not a distinction that can be made in the context of standard survey protocols, so the two are combined for practical purposes.
Aggregate data on religious service attendance, which are the most factual, have a different meaning in countries with little religious pluralism: the obligation to attend places of worship varies between religions and does not apply in the same way to men and women (see Breton 2009).
See the Crombach “alpha” in the appendix (inter-item correlations).
Danièle Hervieux-Léger (1999: 71–9) devotes a chapter to the end of inherited religious identities (as there is scarcely a satisfactory measure of belief, the decoupling relates to low levels of practice even among those who claim a religious identity, and not necessarily the idea that believers and practitioners are separate groups).
I only used the Eurobarometers for a comparison of average values over time.
No country in the Sahel (with the exception of Nigeria, which is partially Sahelian) or in North Africa has been the subject of studies at different dates.
The term ethno-cultural groups is used to designate people from diverse countries in regions such as the Maghreb and Southern Europe which share a common culture and history but not necessarily a common ethnicity.
Regarding men in France, the ISSP asked the question “Do you believe in God?”, and 60% responded: “I don’t” in 1991, and 59% in 2008. The ISSP asked “How religious are you?” and 29% responded “not at all religious” in 2002 and in 2008. Finally, the EVS asked “Do you belong to a religious denomination?”: 39% of men responded positively in 1991 and 54% in 2008. No questions precisely matched the MGIS and TEO formulations but the progression of irreligion that they show are either zero or 15 percentage points, or even lower than suggested by a comparison between MGIS and TEO.
The notion of regular practice correlates empirically with practising once a month according to calculations we have done on the majority population in France (as can be checked with the data in the appendix).
Attendance at Friday prayer is not obligatory for women, slaves, the sick and children, as it is for other people (Qur’an, Surah 62, verse 9). Regarding overt symbols, requirements also differ depending on sex: women are expected to cover their necks and throats (Qur’an, Surah 24, verse 31).
Neither of the two surveys mentioned prayer, which however is included in most international studies.
Asian countries where religious issues are different again—while, like Europe and central Africa, they might have illuminated by contrasting the dynamics observed among immigrants from Muslim countries—are not taken into account here.
The Islamic Revolution in Iran, the resistance to the Soviet intervention in Afghanistan, the assassination of Anwar Sadat and the rise of forces like the Islamic Salvation Front in Algeria, Hamas in Palestine and Hezbollah Lebanon.
See the figures in the appendix.
The word secularization encompasses two dimensions: firstly a disentanglement of the religious from the worldly, next a trend towards a decline in collective religious practices and the ebbing of religiosity in the public sphere towards the private sphere. On secularization and “disenchantment,” see Gauchet (1985) and Taylor (2011, chap. 3). Hereafter, and through taking liberties with the terminology, I will use “secularization” to mean a decline in practice.
Among men over 45 years, around 80% attend collective Friday prayers at least monthly, 60% of those aged 15– 44 years do so.
In Turkey, there is a sizeable core of practice and a pronounced process that sees levels of practice increase with age. But, according to the WVS surveys, the level of practice has hardly increased in recent decades and the coming to power of the AKP, in 2002, did not significantly alter developments already begun.
Norris and Inglehart (2004: 87).
After 1990 there are no data based on a common protocol.
Y. Lambert, “Sortie du religieux, dilution du religieux” in Lambert and Michelat (1992: 40).
Tables 2a and 2b present the percentages of men and women separately.
Those from central Africa and the Gulf of Guinea are distinguishable in the 1992 survey.
One might question the use, in the law, of the word “ostentatoire” (“ostentatious”—which wants to be seen). “Ostensible” (“conspicuous”—which can be seen) might have been less offensive.
Note also that there is little difference between men and women among second generation Europeans.
Given that there were no surveys two decades apart in the African countries under consideration, we cannot accurately objectify this rise.
The EVS clearly show that in the countries with an Orthodox religious majority (Bulgaria, Greece, Serbia, Ukraine), religiosity measured by monthly practice was high for all age groups between 1990 and 2008.
In addition to immigration from the centre and the west of the country there is (Turkish) Kurdish immigration from the east which, on the whole, will be more secular (Tribalat et al. 1996: 241).
Six immigrant groups can be compared— Algeria, Morocco, Turkey, Sahel, central and Guinean Africa, Spain–Portugal—but only two groups of descendants—Algeria and Spain–Portugal.
This latter aspect is in contradiction with the persistence or increase in religiosity among young immigrants from Africa and Turkey. One of the possible explanations of the surprising decline in practice by Sahelian African immigrants is that in 1992 among these men, often married in their home countries, was a large contingent living in hostels (where the intensity of practice is greater than for those living outside of hostels) (Tribalat et al. 1996). Such a situation had become uncommon in 2008.
The reasons that led me to only cautiously compare the percentages of those who declare themselves “without religion” in the two French surveys do not make it impossible to make a comparison between different ethno-cultural groups in the same survey, nor to make comparisons of age profiles, level of education or type of neighbourhoods from one survey to another.
An analysis of three intertwined effects of age, cohort and period would require a series of surveys with the same protocols to calculate coefficients specific to each dimension of the temporal process.
According to MGIS, we can only note that among young people aged under 31 from the second generation of Algerian immigration, there is a trend towards declining practice with age.
We are limited to those under 51 years if we want to compare immigrants and descendants of immigrants.
Attending a religious service does not vary among immigrants from central Africa and the Gulf of Guinea.
Among immigrants who were over 16 years when they arrived, if we omit those from Turkey, practically unchanged percentages were recorded by age.
With immigrants and descendants of immigrants.
In fact, Frank Van Tubergen and Jorunn Sindradottir (2011) reached the conclusion that a standard deviation increase in level of education is associated with a decline in weekly worship attendance: this might perhaps be linked with the weight of different immigrant groups in their analysis.
An aetiology of trends toward secularization has been carried out at the macrosocial scale on a sample of thirty countries (Höllinger and Haller 2009), comparing societies as units of observation based on the WVS and above all the ISSP.
Translator’s note: An IRIS is an “aggregated unit for statistical information,” and a residential IRIS has a population that “generally falls between 1,800 and 5,000. The unit is homogeneous in terms of living environment and the boundaries of the unit are based on the major dividing lines provided by the urban fabric (main roads, railways, bodies of water etc.)” (http://www.insee.fr/en/methodes/default.asp?page=definitions/iris.htm).
In a multi-level analysis on immigrants in Europe, Frank Van Tubergen and Jorunn Sindradottir (2011) observed that religiosity is fairly sensitive to experiencing unemployment.
Translator’s note: A ZUS is a “sensitive urban zone”; an infra-urban area defined by the authorities as being a priority target for urban policy, according to local factors relating to the difficulties that the inhabitants of these areas experience (http://www.insee.fr/en/methodes/default.asp?page=definitions/zone-urbaine-sensible.htm).
See Beauchemin, Hamel, Simon, forth-coming in 2014, “Trajectoires et origines enquête sur la diversité des populations en France” (provisional title), Paris: INED (collection Grandes Enquêtes).
There is, however, a significant impact from segregation on the religiosity of members of the majority group.
It has in fact been observed that this is a sort of background process (see for example Alba and Nee 2003).
We tested a variant with zone of residence either being a ZUS or non-ZUS.
This latter delineation facilitates comparisons with the MGIS survey; the other delineations of arrival periods were established by the TEO team.
In cohorts for which there is a sufficiently wide spectrum of ages.
There is all the more reason for doing this since we do not have data on religiosity for the older age groups.
The models, estimated using the Stata xtmelogit procedure, are based on sub-groups of observations defined by generation, ancestry and arrival in France cohorts for the two most populous groups of origin (European immigrants and immigrants from the Sahel, Maghreb and Turkey). Here we test whether levels of religiosity are identical from one sub-group to another and this being the case, whether the age effect on religiosity varies between these sub-groups.
The high level of deviation from the mean in the cohort that arrived after 1992 could relate to the greater number of immigrants from eastern Europe
For them the religiosity declared in France is an element of continuity with their lives in their countries of origin. For the most part, it is not constructed in opposition to the central currents in their societies of origin: the lower religiosity shown by the oldest immigrant cohorts could reflect the fact that they were socialized at a time when religious vitality was less strong.
We have compared descendants of immigrants aged 36– 46 years in 2008 having been adolescents in France between 1981 and 1991 to descendents of immigrants aged 20–30 years in 1992, and socialized in France in the same period: statistically they are members of the same cohort followed “longitudinally.” Yet, among the eldest descendants of Algerian immigrants (the first cited are those aged 20–30 years and the second those aged 36– 46 years), claiming a religion is sustained with age (2 immigrant parents: 70–68%, 1 parent 45– 44%) but practice declines (2 parents: 14–7%, 1 parent 6–3%). While for descendants of immigrants from Spain and Portugal, the proportion claiming a religious identity is much lower for the more elderly (2 parents: 77–68%, 1 parent: 64–53%), and the level of practice, which is more or less equal is however very low (2 parents: 6–7% and 1 parent: 3– 4% respectively go to mass once a month). Can we hypothesize that time spent living in France dilutes the religiosity of the 36– 46-year-olds coming from Spain and Portugal, but only has a weak impact on descendants of Algerian immigrants?
We do not yet have sufficient data on the second generation whose parents arrived after 1974, but their level of religiosity does not decline as they advance in age.
It is not a question of explaining religious attitudes, as it is with objectified characteristics, but of a rationale of understanding. The assertion of an identity such as “I am Maghrebi” is on the same plane, it can make a sense of religiosity explicit but does not pretend to explain it. Adherence to the Ummah has not been measured as such.
While there is a marked mistrust of the police, there is not however a general distrust of other institutions.
There are also identity assertions that are not, or are only indirectly, instrumental, and that bring about symbolic rewards for those who adopt them without providing advantages in social and economic competition.
Access to employment and housing depends not only on attitudes and personal investments.
Thus, the worship of domestic gods and sacrifices continued for a long time in the post-Constantine Roman Empire; the Marranos formed themselves into and asserted themselves as a distinct identity during the Spanish Inquisition, and the behaviour of Catholics in Ulster is testament to the impact of hostility, suspicion and repression of minority identities. The maintenance of minority identities, which are often hidden because of repressive social norms and despite massacres and persecution, has sometimes been open, as shown by ancient pluralistic contexts such as the Ottoman Empire with its Millet system.
This distinction is not the same as the distinction between declarative and non-declarative. Without exception, with this survey data we do not have observations of practices such as counts an observer could make by going to a parish. But a respondent’s statements do not have the same status all the same. We are led to distinguish between the statements that do not relate to a practice (propositional values) from those that target the same subjects (such as a count of parishioners for example), or though a subject’s declaration such as: “I go to the parish church every Sunday.”